Traveling Salesperson Problem on TSP 40 nodes
2.2Optimality GapGNARLBC
Evaluation Results
| Method | Links | |
|---|---|---|
| GNARLBCtraining=Behavioral Cloning2025.09 | 2.2 | |
| GNARLPPOtraining=PPO2025.09 | 3 | |
| GNARLPPO+heurtraining=PPO, insertion_heuristic=true2025.09 | 5.3 | |
| Georgiev et al.beam_search_width=12802025.09 | 7.5 | |
| Christofidestype=handcrafted heuristic2025.09 | 10.1 |